cmdline: add collect-logs subcommand.

Add a new collect-logs sub command to the cloud-init CLI. This script
will collect all logs pertinent to a cloud-init run and store them in a
compressed tar-gzipped file. This tarfile can be attached to any
cloud-init bug filed in order to aid in bug triage and resolution.

A cloudinit.apport module is also added that allows apport interaction.
Here is an example bug filed via ubuntu-bug cloud-init: LP: #1716975.

Once the apport launcher is packaged in cloud-init, bugs can be filed
against cloud-init with the following command:
  ubuntu-bug cloud-init

LP: #1607345

5 files changed, 336 insertions, 1 deletions

diff --git a/cloudinit/apport.py b/cloudinit/apport.py
new file mode 100644
index 00000000..221f341c
--- /dev/null
+++ b/cloudinit/apport.py
@@ -0,0 +1,105 @@
+# Copyright (C) 2017 Canonical Ltd.
+#
+# This file is part of cloud-init. See LICENSE file for license information.
+
+'''Cloud-init apport interface'''
+
+try:
+    from apport.hookutils import (
+        attach_file, attach_root_command_outputs, root_command_output)
+    has_apport = True
+except ImportError:
+    has_apport = False
+
+
+KNOWN_CLOUD_NAMES = [
+    'Amazon - Ec2', 'AliYun', 'AltCloud', 'Azure', 'Bigstep', 'CloudSigma',
+    'CloudStack', 'DigitalOcean', 'GCE - Google Compute Engine', 'MAAS',
+    'NoCloud', 'OpenNebula', 'OpenStack', 'OVF', 'Scaleway', 'SmartOS',
+    'VMware', 'Other']
+
+# Potentially clear text collected logs
+CLOUDINIT_LOG = '/var/log/cloud-init.log'
+CLOUDINIT_OUTPUT_LOG = '/var/log/cloud-init-output.log'
+USER_DATA_FILE = '/var/lib/cloud/instance/user-data.txt'  # Optional
+
+
+def attach_cloud_init_logs(report, ui=None):
+    '''Attach cloud-init logs and tarfile from 'cloud-init collect-logs'.'''
+    attach_root_command_outputs(report, {
+        'cloud-init-log-warnings':
+            'egrep -i "warn|error" /var/log/cloud-init.log',
+        'cloud-init-output.log.txt': 'cat /var/log/cloud-init-output.log'})
+    root_command_output(
+        ['cloud-init', 'collect-logs', '-t', '/tmp/cloud-init-logs.tgz'])
+    attach_file(report, '/tmp/cloud-init-logs.tgz', 'logs.tgz')
+
+
+def attach_hwinfo(report, ui=None):
+    '''Optionally attach hardware info from lshw.'''
+    prompt = (
+        'Your device details (lshw) may be useful to developers when'
+        ' addressing this bug, but gathering it requires admin privileges.'
+        ' Would you like to include this info?')
+    if ui and ui.yesno(prompt):
+        attach_root_command_outputs(report, {'lshw.txt': 'lshw'})
+
+
+def attach_cloud_info(report, ui=None):
+    '''Prompt for cloud details if available.'''
+    if ui:
+        prompt = 'Is this machine running in a cloud environment?'
+        response = ui.yesno(prompt)
+        if response is None:
+            raise StopIteration  # User cancelled
+        if response:
+            prompt = ('Please select the cloud vendor or environment in which'
+                      ' this instance is running')
+            response = ui.choice(prompt, KNOWN_CLOUD_NAMES)
+            if response:
+                report['CloudName'] = KNOWN_CLOUD_NAMES[response[0]]
+            else:
+                report['CloudName'] = 'None'
+
+
+def attach_user_data(report, ui=None):
+    '''Optionally provide user-data if desired.'''
+    if ui:
+        prompt = (
+            'Your user-data or cloud-config file can optionally be provided'
+            ' from {0} and could be useful to developers when addressing this'
+            ' bug. Do you wish to attach user-data to this bug?'.format(
+                USER_DATA_FILE))
+        response = ui.yesno(prompt)
+        if response is None:
+            raise StopIteration  # User cancelled
+        if response:
+            attach_file(report, USER_DATA_FILE, 'user_data.txt')
+
+
+def add_bug_tags(report):
+    '''Add any appropriate tags to the bug.'''
+    if 'JournalErrors' in report.keys():
+        errors = report['JournalErrors']
+        if 'Breaking ordering cycle' in errors:
+            report['Tags'] = 'systemd-ordering'
+
+
+def add_info(report, ui):
+    '''This is an entry point to run cloud-init's apport functionality.
+
+    Distros which want apport support will have a cloud-init package-hook at
+    /usr/share/apport/package-hooks/cloud-init.py which defines an add_info
+    function and returns the result of cloudinit.apport.add_info(report, ui).
+    '''
+    if not has_apport:
+        raise RuntimeError(
+            'No apport imports discovered. Apport functionality disabled')
+    attach_cloud_init_logs(report, ui)
+    attach_hwinfo(report, ui)
+    attach_cloud_info(report, ui)
+    attach_user_data(report, ui)
+    add_bug_tags(report)
+    return True
+
+# vi: ts=4 expandtab
diff --git a/cloudinit/cmd/devel/logs.py b/cloudinit/cmd/devel/logs.py
new file mode 100644
index 00000000..35ca478f
--- /dev/null
+++ b/cloudinit/cmd/devel/logs.py
@@ -0,0 +1,101 @@
+# Copyright (C) 2017 Canonical Ltd.
+#
+# This file is part of cloud-init. See LICENSE file for license information.
+
+"""Define 'collect-logs' utility and handler to include in cloud-init cmd."""
+
+import argparse
+from cloudinit.util import (
+    ProcessExecutionError, chdir, copy, ensure_dir, subp, write_file)
+from cloudinit.temp_utils import tempdir
+from datetime import datetime
+import os
+import shutil
+
+
+CLOUDINIT_LOGS = ['/var/log/cloud-init.log', '/var/log/cloud-init-output.log']
+CLOUDINIT_RUN_DIR = '/run/cloud-init'
+USER_DATA_FILE = '/var/lib/cloud/instance/user-data.txt'  # Optional
+
+
+def get_parser(parser=None):
+    """Build or extend and arg parser for collect-logs utility.
+
+    @param parser: Optional existing ArgumentParser instance representing the
+        collect-logs subcommand which will be extended to support the args of
+        this utility.
+
+    @returns: ArgumentParser with proper argument configuration.
+    """
+    if not parser:
+        parser = argparse.ArgumentParser(
+            prog='collect-logs',
+            description='Collect and tar all cloud-init debug info')
+    parser.add_argument(
+        "--tarfile", '-t', default='cloud-init.tar.gz',
+        help=('The tarfile to create containing all collected logs.'
+              ' Default: cloud-init.tar.gz'))
+    parser.add_argument(
+        "--include-userdata", '-u', default=False, action='store_true',
+        dest='userdata', help=(
+            'Optionally include user-data from {0} which could contain'
+            ' sensitive information.'.format(USER_DATA_FILE)))
+    return parser
+
+
+def _write_command_output_to_file(cmd, filename):
+    """Helper which runs a command and writes output or error to filename."""
+    try:
+        out, _ = subp(cmd)
+    except ProcessExecutionError as e:
+        write_file(filename, str(e))
+    else:
+        write_file(filename, out)
+
+
+def collect_logs(tarfile, include_userdata):
+    """Collect all cloud-init logs and tar them up into the provided tarfile.
+
+    @param tarfile: The path of the tar-gzipped file to create.
+    @param include_userdata: Boolean, true means include user-data.
+    """
+    tarfile = os.path.abspath(tarfile)
+    date = datetime.utcnow().date().strftime('%Y-%m-%d')
+    log_dir = 'cloud-init-logs-{0}'.format(date)
+    with tempdir(dir='/tmp') as tmp_dir:
+        log_dir = os.path.join(tmp_dir, log_dir)
+        _write_command_output_to_file(
+            ['dpkg-query', '--show', "-f=${Version}\n", 'cloud-init'],
+            os.path.join(log_dir, 'version'))
+        _write_command_output_to_file(
+            ['dmesg'], os.path.join(log_dir, 'dmesg.txt'))
+        _write_command_output_to_file(
+            ['journalctl', '-o', 'short-precise'],
+            os.path.join(log_dir, 'journal.txt'))
+        for log in CLOUDINIT_LOGS:
+            copy(log, log_dir)
+        if include_userdata:
+            copy(USER_DATA_FILE, log_dir)
+        run_dir = os.path.join(log_dir, 'run')
+        ensure_dir(run_dir)
+        shutil.copytree(CLOUDINIT_RUN_DIR, os.path.join(run_dir, 'cloud-init'))
+        with chdir(tmp_dir):
+            subp(['tar', 'czvf', tarfile, log_dir.replace(tmp_dir + '/', '')])
+
+
+def handle_collect_logs_args(name, args):
+    """Handle calls to 'cloud-init collect-logs' as a subcommand."""
+    collect_logs(args.tarfile, args.userdata)
+
+
+def main():
+    """Tool to collect and tar all cloud-init related logs."""
+    parser = get_parser()
+    handle_collect_logs_args('collect-logs', parser.parse_args())
+    return 0
+
+
+if __name__ == '__main__':
+    main()
+
+# vi: ts=4 expandtab
diff --git a/cloudinit/cmd/devel/tests/__init__.py b/cloudinit/cmd/devel/tests/__init__.py
new file mode 100644
index 00000000..e69de29b
--- /dev/null
+++ b/cloudinit/cmd/devel/tests/__init__.py
diff --git a/cloudinit/cmd/devel/tests/test_logs.py b/cloudinit/cmd/devel/tests/test_logs.py
new file mode 100644
index 00000000..dc4947cc
--- /dev/null
+++ b/cloudinit/cmd/devel/tests/test_logs.py
@@ -0,0 +1,120 @@
+# This file is part of cloud-init. See LICENSE file for license information.
+
+from cloudinit.cmd.devel import logs
+from cloudinit.util import ensure_dir, load_file, subp, write_file
+from cloudinit.tests.helpers import FilesystemMockingTestCase, wrap_and_call
+from datetime import datetime
+import os
+
+
+class TestCollectLogs(FilesystemMockingTestCase):
+
+    def setUp(self):
+        super(TestCollectLogs, self).setUp()
+        self.new_root = self.tmp_dir()
+        self.run_dir = self.tmp_path('run', self.new_root)
+
+    def test_collect_logs_creates_tarfile(self):
+        """collect-logs creates a tarfile with all related cloud-init info."""
+        log1 = self.tmp_path('cloud-init.log', self.new_root)
+        write_file(log1, 'cloud-init-log')
+        log2 = self.tmp_path('cloud-init-output.log', self.new_root)
+        write_file(log2, 'cloud-init-output-log')
+        ensure_dir(self.run_dir)
+        write_file(self.tmp_path('results.json', self.run_dir), 'results')
+        output_tarfile = self.tmp_path('logs.tgz')
+
+        date = datetime.utcnow().date().strftime('%Y-%m-%d')
+        date_logdir = 'cloud-init-logs-{0}'.format(date)
+
+        expected_subp = {
+            ('dpkg-query', '--show', "-f=${Version}\n", 'cloud-init'):
+                '0.7fake\n',
+            ('dmesg',): 'dmesg-out\n',
+            ('journalctl', '-o', 'short-precise'): 'journal-out\n',
+            ('tar', 'czvf', output_tarfile, date_logdir): ''
+        }
+
+        def fake_subp(cmd):
+            cmd_tuple = tuple(cmd)
+            if cmd_tuple not in expected_subp:
+                raise AssertionError(
+                    'Unexpected command provided to subp: {0}'.format(cmd))
+            if cmd == ['tar', 'czvf', output_tarfile, date_logdir]:
+                subp(cmd)  # Pass through tar cmd so we can check output
+            return expected_subp[cmd_tuple], ''
+
+        wrap_and_call(
+            'cloudinit.cmd.devel.logs',
+            {'subp': {'side_effect': fake_subp},
+             'CLOUDINIT_LOGS': {'new': [log1, log2]},
+             'CLOUDINIT_RUN_DIR': {'new': self.run_dir}},
+            logs.collect_logs, output_tarfile, include_userdata=False)
+        # unpack the tarfile and check file contents
+        subp(['tar', 'zxvf', output_tarfile, '-C', self.new_root])
+        out_logdir = self.tmp_path(date_logdir, self.new_root)
+        self.assertEqual(
+            '0.7fake\n',
+            load_file(os.path.join(out_logdir, 'version')))
+        self.assertEqual(
+            'cloud-init-log',
+            load_file(os.path.join(out_logdir, 'cloud-init.log')))
+        self.assertEqual(
+            'cloud-init-output-log',
+            load_file(os.path.join(out_logdir, 'cloud-init-output.log')))
+        self.assertEqual(
+            'dmesg-out\n',
+            load_file(os.path.join(out_logdir, 'dmesg.txt')))
+        self.assertEqual(
+            'journal-out\n',
+            load_file(os.path.join(out_logdir, 'journal.txt')))
+        self.assertEqual(
+            'results',
+            load_file(
+                os.path.join(out_logdir, 'run', 'cloud-init', 'results.json')))
+
+    def test_collect_logs_includes_optional_userdata(self):
+        """collect-logs include userdata when --include-userdata is set."""
+        log1 = self.tmp_path('cloud-init.log', self.new_root)
+        write_file(log1, 'cloud-init-log')
+        log2 = self.tmp_path('cloud-init-output.log', self.new_root)
+        write_file(log2, 'cloud-init-output-log')
+        userdata = self.tmp_path('user-data.txt', self.new_root)
+        write_file(userdata, 'user-data')
+        ensure_dir(self.run_dir)
+        write_file(self.tmp_path('results.json', self.run_dir), 'results')
+        output_tarfile = self.tmp_path('logs.tgz')
+
+        date = datetime.utcnow().date().strftime('%Y-%m-%d')
+        date_logdir = 'cloud-init-logs-{0}'.format(date)
+
+        expected_subp = {
+            ('dpkg-query', '--show', "-f=${Version}\n", 'cloud-init'):
+                '0.7fake',
+            ('dmesg',): 'dmesg-out\n',
+            ('journalctl', '-o', 'short-precise'): 'journal-out\n',
+            ('tar', 'czvf', output_tarfile, date_logdir): ''
+        }
+
+        def fake_subp(cmd):
+            cmd_tuple = tuple(cmd)
+            if cmd_tuple not in expected_subp:
+                raise AssertionError(
+                    'Unexpected command provided to subp: {0}'.format(cmd))
+            if cmd == ['tar', 'czvf', output_tarfile, date_logdir]:
+                subp(cmd)  # Pass through tar cmd so we can check output
+            return expected_subp[cmd_tuple], ''
+
+        wrap_and_call(
+            'cloudinit.cmd.devel.logs',
+            {'subp': {'side_effect': fake_subp},
+             'CLOUDINIT_LOGS': {'new': [log1, log2]},
+             'CLOUDINIT_RUN_DIR': {'new': self.run_dir},
+             'USER_DATA_FILE': {'new': userdata}},
+            logs.collect_logs, output_tarfile, include_userdata=True)
+        # unpack the tarfile and check file contents
+        subp(['tar', 'zxvf', output_tarfile, '-C', self.new_root])
+        out_logdir = self.tmp_path(date_logdir, self.new_root)
+        self.assertEqual(
+            'user-data',
+            load_file(os.path.join(out_logdir, 'user-data.txt')))
diff --git a/cloudinit/cmd/main.py b/cloudinit/cmd/main.py
index 68563e0c..6fb9d9e7 100644
--- a/cloudinit/cmd/main.py
+++ b/cloudinit/cmd/main.py
@@ -764,16 +764,25 @@ def main(sysv_args=None):
     parser_devel = subparsers.add_parser(
         'devel', help='Run development tools')
 
+    parser_collect_logs = subparsers.add_parser(
+        'collect-logs', help='Collect and tar all cloud-init debug info')
+
     if sysv_args:
         # Only load subparsers if subcommand is specified to avoid load cost
         if sysv_args[0] == 'analyze':
             from cloudinit.analyze.__main__ import get_parser as analyze_parser
             # Construct analyze subcommand parser
             analyze_parser(parser_analyze)
-        if sysv_args[0] == 'devel':
+        elif sysv_args[0] == 'devel':
             from cloudinit.cmd.devel.parser import get_parser as devel_parser
             # Construct devel subcommand parser
             devel_parser(parser_devel)
+        elif sysv_args[0] == 'collect-logs':
+            from cloudinit.cmd.devel.logs import (
+                get_parser as logs_parser, handle_collect_logs_args)
+            logs_parser(parser_collect_logs)
+            parser_collect_logs.set_defaults(
+                action=('collect-logs', handle_collect_logs_args))
 
     args = parser.parse_args(args=sysv_args)